By Matt Harrington, BCBA · Behaviorist Book Club · April 2026 · 12 min read

PECS Implementation in ABA: A Phase-by-Phase Guide for Behavior Analysts

Overview & Clinical Significance

The Picture Exchange Communication System (PECS) is a structured augmentative and alternative communication (AAC) program designed to teach functional communication to individuals who lack reliable speech or whose verbal communication is insufficient to meet their daily needs. Developed by Bondy and Frost in the context of the Delaware Autism Program, PECS teaches individuals to initiate communication by handing a picture to a communicative partner in exchange for a desired item or activity. The system progresses through six sequential phases, each building on the communication skills established in the previous phase.

PECS is grounded in applied behavior analysis. Its teaching methodology draws on Skinner's verbal behavior framework, specifically the mand relation: a verbal operant whose form is controlled by a motivating operation and that is reinforced by the specific object or event specified. By teaching individuals to exchange pictures as a form of manding, PECS establishes a functional communicative repertoire even in the absence of speech. Critically, PECS begins with communication that is inherently motivating — requesting preferred items — rather than beginning with imitation or labeling tasks that may not have functional reinforcement for the learner.

For behavior analysts, PECS is notable for its alignment with ABA principles across all phases of instruction. Phase I uses physical guidance and systematic prompt fading to teach the physical exchange. Phase II extends the exchange across distance. Phase III teaches discrimination between pictures. Phase IV introduces sentence structure through the use of an 'I want' card. Phase V teaches responding to 'What do you want?' And Phase VI adds commenting, extending the communicative repertoire beyond manding.

The empirical literature consistently supports PECS as an evidence-based practice for individuals with autism spectrum disorder and related developmental disabilities. Meta-analyses and systematic reviews identify PECS as effective for establishing functional communication, with additional benefits including reductions in problem behavior maintained by escape from communicative demands and increases in spontaneous speech in some learners.

Background & Context

Prior to the development of PECS, many AAC approaches for individuals with significant communication disabilities required fine motor skills (sign language) or technological access (voice output devices) that posed barriers for some learners. PECS was specifically designed to be accessible to individuals with limited motor skills, limited imitative repertoires, and significant cognitive or language delays — the learner characteristics most commonly associated with autism spectrum disorder and related disabilities.

Skinner's verbal behavior analysis provides the theoretical scaffolding for understanding why PECS works. The mand is the first verbal operant targeted in PECS because it is the verbal behavior that most directly benefits the speaker — the learner requests what they want and receives it, creating an immediate and natural reinforcement consequence. This is in contrast to tact or intraverbal training, which may require the learner to respond to an adult's agenda rather than their own motivating operations.

The two-instructor format used in PECS Phase I is a deliberate procedural design that separates the communicative partner from the physical prompter. The communicative partner holds the preferred item and waits for the exchange; the physical prompter, positioned behind the learner, uses hand-over-hand guidance to physically prompt the exchange without speaking. This arrangement ensures that the social contingency — the exchange — is never contaminated with a physical prompt from the person receiving the picture, which would create a prompt dependency to the communicative partner's physical presence.

Generalization of PECS across communicative partners, settings, and picture stimuli must be programmed explicitly. Learners who can exchange a single picture with a familiar therapist in a clinic setting have not demonstrated functional communication — they have demonstrated a behavior under a specific set of stimulus conditions. Generalization programming involves training with multiple partners and in multiple settings, using varied picture stimuli across a range of motivating conditions.

Speech-generating devices and high-tech AAC have become increasingly available and accessible, and BCBAs must be prepared to evaluate whether PECS, low-tech picture systems, or speech-generating devices are most appropriate for a given learner — either independently or as a progression. PECS is often used as a stepping stone to high-tech AAC and may facilitate the development of speech in some learners.

Clinical Implications

PECS Phase I instruction requires the two-instructor format, and BCBAs must train both roles — the communicative partner and the physical prompter — before implementation begins. Common implementation errors include: the communicative partner prompting the learner verbally or physically during the exchange, the prompter failing to fade physical guidance systematically, and delivering reinforcement with a significant delay between the exchange and access to the preferred item. Each of these errors compromises the clarity of the contingency and slows acquisition.

Motivating operations are the engine of PECS instruction. Instruction should occur when the learner's motivating operation for the target item or activity is at its highest — before access to the item, at the beginning of preferred activity periods, during naturally motivating contexts such as snack time or play. Attempting PECS instruction during periods of satiation, after extensive access to preferred items, or during non-preferred activities significantly reduces learning efficiency. Pre-session preference assessments confirm that the items used in instruction are currently functioning as reinforcers.

Phase III discrimination training is frequently where learners plateau if procedures are not implemented carefully. Teaching picture discrimination requires systematic use of errorless and error-correction procedures. A common approach begins by teaching discrimination between the target picture and a clearly irrelevant foil (e.g., a blank card), then progressively introducing increasingly similar foils. When a learner exchanges an incorrect picture, the error correction procedure — typically four-step: prevent access, redirect, prompt correct exchange, reinforce — must be implemented consistently and without emotional expression to avoid inadvertent reinforcement of the error.

PECS Phase IV, which introduces the sentence strip and 'I want' card, requires that the learner already has a robust picture exchange repertoire. Attempting Phase IV before Phase III mastery produces confusion about the communicative format and typically requires returning to earlier phases. Mastery criteria for each phase — including generalization criteria — should be specified in writing before the program begins and applied consistently across implementers.

Ethical Considerations

BACB Ethics Code 2.01 requires behavior analysts to provide scientifically supported treatment. PECS has extensive empirical support as an evidence-based practice for functional communication in individuals with autism spectrum disorder, and its alignment with ABA methodology makes it an appropriate choice for behavior analysts supporting learners with significant communication deficits. BCBAs who implement PECS are applying evidence-based practice; those who fail to address functional communication in learners who cannot reliably communicate their needs are at risk of violating Code 2.09 by omitting a clinically essential intervention.

Code 2.09 also addresses the use of least-intrusive procedures. PECS instruction, particularly in early phases, involves physical guidance — a form of physical prompting that requires learner physical contact. BCBAs must ensure that informed consent is obtained for physical guidance procedures, that prompting is delivered with the minimum intrusiveness necessary, and that physical guidance is faded systematically as the learner acquires the exchange behavior. The use of physical guidance should be documented in the behavior intervention plan.

Access to functional communication is a fundamental right, and BCBAs have an ethical obligation to advocate for communication programming when learners lack reliable communication systems. Code 1.01 addresses beneficence — BCBAs must act in the best interest of clients — and denying or delaying functional communication programming to a non-vocal learner fails to meet this standard. BCBAs should be prepared to advocate for PECS or appropriate AAC services even in the face of administrative, financial, or philosophical barriers.

Cultural considerations are relevant to PECS implementation. The pictures used in PECS books and boards should reflect the learner's cultural context, family environment, and preferred activities. Using pictures that are culturally unfamiliar or that do not represent the learner's actual reinforcers reduces the functional relevance of the system and compromises its effectiveness. Families should be involved in selecting pictures and in understanding and implementing the system at home.

Assessment & Decision-Making

Candidacy assessment for PECS begins with a functional communication assessment: What communication repertoire does the learner currently have? Can they reliably request preferred items through any modality — pointing, reaching, vocalizing, signing? What is their reinforcer profile? Do they approach communicative partners or avoid them? Answers to these questions guide the decision about whether PECS is the appropriate starting point or whether prerequisite skills — approach behavior, attending, or basic motor skills — need to be established first.

A preference assessment conducted immediately before each PECS session confirms that the items used are currently functioning as reinforcers. Learners who are satiated on the instructional item will not be motivated to exchange pictures for access to it. Using probe trials — offering the item non-contingently and observing whether the learner reaches for it — provides a quick check of motivating operation strength before the session begins.

Progress monitoring in PECS should track phase-level performance, including: number of exchanges per session, percentage of prompted versus independent exchanges, accuracy on Phase III discrimination, sentence strip length in Phase IV, and spontaneous versus prompted initiation. These data support decisions about when to advance to the next phase, when to add new pictures to the learner's book, and when to conduct generalization probes across partners and settings.

When a learner plateaus within a PECS phase, the BCBA should systematically evaluate: the fidelity of implementation across all instructors, the potency of the reinforcers being used, the adequacy of the prompting and fading procedure, and whether prerequisite skills for that phase are solidly established. Phase III errors, for example, may reflect insufficient stimulus control training rather than a learner deficit — adjusting the errorless learning procedures often resolves the plateau.

Decision-making about transitioning from PECS to high-tech AAC should be individualized. Some learners progress through PECS phases and transition naturally to speech-generating devices; others use PECS long-term as their primary communication system. The decision should be driven by the learner's communication needs, family preferences, and empirical data on device use — not by a predetermined assumption that technology is always preferable to low-tech systems.

What This Means for Your Practice

If you work with non-vocal or minimally verbal learners, competency in PECS implementation is a clinical necessity. This means knowing not only the six phases but the specific instructional procedures for each — the two-instructor format in Phase I, the errorless discrimination procedures in Phase III, the sentence strip protocol in Phase IV — and being able to teach these procedures with fidelity to the RBTs and ABATs you supervise.

Training staff in PECS requires hands-on practice. Reading about PECS or watching video demonstrations does not produce implementation fidelity. Staff need to practice the physical prompter role, experience the timing demands of the communicative partner role, and receive specific feedback on error correction delivery. Structured role-plays, with the supervisor playing the learner, are effective training methods that can be conducted within supervision sessions.

Family involvement is essential and should begin before PECS instruction starts. Parents who understand the phase structure, who know what their child is working on and why, and who can implement PECS at home during naturally motivating contexts — meal times, snack time, play — extend the impact of clinic-based instruction dramatically. Provide families with a simple phase summary, teach them the two-instructor format, and practice with them directly.

Collaborate with SLPs when implementing PECS. Speech-language pathologists bring expertise in communication development, AAC assessment, and augmentative communication device selection that complements the behavior analyst's instructional expertise. Joint programming — where the SLP contributes to picture selection, sentence structure goals, and generalization planning while the BCBA manages reinforcement contingencies and prompting procedures — produces better outcomes than either discipline working in isolation.

Finally, keep generalization at the forefront of every PECS program. A learner who exchanges pictures only with their ABA therapist in the clinic has not achieved functional communication. Build generalization training into every phase, track generalization data as part of your standard progress monitoring, and do not consider a phase mastered until the learner can perform it across multiple partners, settings, and picture stimuli under naturalistic motivating conditions.